The Multiple Risk Factor Intervention Trial (MRFIT) revealed an unexpected subgroup finding: an association between diuretic treatment (especially with hydrochlorothiazide) and an increased rate of sudden death in hypertensive men with LVH and other ECG abnormalities. The question is; does this finding result from random variation, or does it represent a serious toxic response to hydrochlorothiazide? Our study is an effort to distinguish between these two possibilities by examining the pathophysiologic mechanism. We will compare the frequency and complexity of ventricular arrhythmias measured blindly on continuous 24 hour electrocardiographic monitoring (CEM) in 400 men aged 35-70 who have the MRFIT-type ECG abnormalities and hypertension, and who have been treated for at least 6 months by their usual physician with either (1) hydrochlorothiazide, (2) hydrochlorothiazide and oral potassium supplementation, (3) hydrochlorothiazide and triamterene, (4) chlorthalidone or (5) no drugs. Serum and red blood cell potassium and magnesium levels will be measured blindly and correlated with ventricular arrhythmias. After this initial cross-sectional comparison of the long term effects of treatment, the men will be withdrawn from all diuretics for one month and then randomized to five treatment regimens (the same as those noted above, with a placebo for the untreated group). Arrhythmias and electrolytes will again be measured blindly after two months of treatment. Echocardiographic measurements of cardiac dimensions and function will be analysed in relation to the ventricular arrhythmias on the different treatment regimens. Approximately 60 million people in the United States have hypertension, and the most commonly prescribed drug is hydrochlorothiazide. The unexpected MRFIT findings raise the serious possibility that more than a quarter of the patients who take this drug-those with ECG abnormalities--may be increasing their risk of sudden death. The size, scope, and design of this study should permit clear conclusions of great public health importance; does hydrochlorothiazide produce the arrhythmias that may predispose such an individual to sudden death, and are chlorthalidone and two potassium-sparing combinations less likely to cause this adverse effect?